Igd Class Switch Recombination Is Not Controlled Through the Immunoglobulin Heavy Chain 3&Prime

LETTER TO THE EDITOR

IgD class switch recombination is not controlled through the immunoglobulin heavy chain 3′ regulatory region super-enhancer

Hussein Issaoui1, Nour Ghazzaui1, Alexis Saintamand2, Yves Denizot and François Boyer

Cellular & Molecular Immunology (2017) 14, 871–874; doi:10.1038/cmi.2017.81; published online 4 September 2017

n secondary lymphoid organs, mature enigmatic event restricted to a few B-cell 3’RR-deficient mouse B cells were used IB cells express membrane immuno- subsets in specific lymphoid tissues (such for these experiments. In some experi- globulin (Ig) of M and D isotypes (IgM as mesenteric lymph nodes, peritoneal ments, 3′RR-deficient mice were and IgD, respectively) of the same spe- cavity and mucosa-associated tissue) in pristane-treated (1 ml) for 2 months to cificity through alternative splicing of a both mice and humans.3–5 A recent induce inflammation prior to the recov- pre-mRNA encompassing the VDJ vari- study suggested that IgD CSR is initiated ery of peritoneal cavity cells. As pre- 5 4 able region and Cμ and Cδ heavy chain by microbiota, demonstrating a role for viously described in detail, junctions constant exons.1 After encountering anti- IgD in the homeostatic regulation of the were amplified using touchdown PCR gen, B cells undergo class switch recom- microbial community. The mechanistic followed by nested PCR. Libraries of bination (CSR) by which the Cμ gene is regulation of IgD CSR remains enig- 200 bp were prepared from the 1–2kb substituted with Cγ, Cε or Cα, thereby matic, reflecting the difficulty to obtain PCR products of Sμ–σδ amplification for generating IgG, IgE and IgA antibodies asufficient number of Sμ–σδ (σ for ion proton sequencing (‘GénoLim plat- of the same antigenic specificity but with S-like) junction sequences for molecular form’ of the Limoges University, France). new effector functions. CSR requires the analysis. We recently reported a new Sequenced reads were subsequently DNA-editing enzyme activation-induced computational tool (CSReport) for the mapped to Sμ and σδ regions using the deaminase (AID), which targets switch automatic analysis of CSR junctions BLAST algorithm. The computational (S) regions preceding Cμ (namely, the Sμ sequenced using high-throughput tool developed for experiments performs 6 fi donor region) and Cγ, Cε and Cα genes sequencing, and we used this tool to junction assembly, identi es break points 1 –σ wt σ fi (namely, the Sγ,ε,α acceptor regions). analyze Sμ δ junctions in wild-type ( ) in Sμ and δ,identies junction struc- CSR is controlled in trans through a mice.7 Genomic deletion of the 3′RR tures (blunt, micro-homology or junc- wide spectrum of enzymes and proteins super-enhancer abrogates CSR to IgG, tions with insertions) and outputs a 2,8 and in cis through the immunoglobulin IgE and IgA classes. In contrast, a statistical summary of the identified heavy chain (IgH) 3′ regulatory region unique study, based on the analysis of a junctions. (3′RR) super-enhancer that specifically few Sμ–σδ junctions, reported the 3′RR In 3′RR-deficient mice, among 4 904 AID-poises the S acceptor but not the Sμ super-enhancer as dispensable for IgD 163 raw reads, 39 046 reads were 4 donor region.2 IgD CSR is a rare and CSR. We thus used CSReport and high- mapped to a switch junction, and the throughput sequencing to conduct a presence of 190 unique Sμ–σδ junctions UMR CNRS 7276, University of Limoges, CBRS, more in depth analysis of Sμ–σδ junc- was documented. The number of unique 87025 Limoges, France. tions in 3′RR-deficient mice. junctions (that is, clonally independent) 1These authors contributed equally to this work. 2Present address: INSERM U1236, Université The present study was approved by was indeed dependent on the biological Rennes 1, Rennes, France. our local ethics committee review board frequency of this CSR event in the Correspondence: Dr Y Denizot, PhD, UMR CNRS (Comité Régional d'Ethique sur l'Expéri- studied samples. Compared to previous 7276, University of Limoges, CBRS, Rue Pr. mentation Animale du Limousin, studies of other IgD7 and IgG6 junctions Descottes, 87025 Limoges, France. wt E-mail: [email protected] Limoges, France) and conducted accord- in animals, IgD CSR clearly appears as Received: 13 July 2017; Revised: 17 July 2017; ing to the European guidelines for ani- a rare event. The structural profiles of Accepted: 18 July 2017 mal experimentation. Peritoneal cavity IgD junctions (blunt, micro-homology 3'RR independent IgD CSR HIssaouiet al

872

or junction with insertions) in 3′RR- The frequency of DNA breaks in Sμ was are reported in Figure 1c (left panel). deficient mice are reported in Figure 1a significantly elevated in AGCT sites com- The breaks detected in the σδ region did (left panel). The positions of IgD junc- pared to the frequency of theoretical not show a significant AID signature, tions in terms of distance from the random breaks. The positions of IgD with most of the break points observed forward PCR primer in Sμ and the junctions in terms of distance from the outside AID motifs. Pristane injection localization of IgD junctions with reverse PCR primer in σδ and the increased the frequency of IgH/c-myc WRCY, RGYW, AGCT and GGG motifs localizations of IgD junctions with translocations; most of the IgH breaks are reported in Figure 1b (left panel). WRCY, RGYW, AGCT and GGG motifs occurred in the Sμ region. We thus used

Cellular & Molecular Immunology 3'RR independent IgD CSR HIssaouiet al

873

Figure 1 Quantitative description of Sμ–σδ junctions using CSReport workflow. Left panel: results from pooled peritoneal B cells obtained from 30 3′RR-deficient mice. Right panel: results from pooled peritoneal B cells from seven 3′RR-deficient mice treated with 1 ml of pristane for 2 months. (a) Structure profiles of Sμ–σδ junctions. Junctions are classified in terms of junction types (junction with insertions, blunt junction or junction with micro-homology) and size (in bp of insertions or micro-homologies). The number in parentheses indicates the number of junctions of each type. The junction profile is significantly different (P = 0.0005, Fisher’s exact test) under basal and pristane-induced inflammatory conditions. (b) Break point localization and motif targeting of Sμ–σδ junctions in the Sμ region (up to 1 kb from the 5′ primer). Arrows indicate the break point positions. Color bars represent the occurrences of relevant AID motifs (forward/reverse hotspots, WRCY (red)/RGYW (blue), palindromic switch hotspot AGCT (green) and cold spot GGG (magenta)). The location of the break points with these motifs is emphasized with a colored asterisk. Proportions of the identified break points on these motifs are represented in the bar graphs (same colors). Significance was assessed based on comparisons with random break positions. For each Sμ–σδ junction data set, N random positions were numerically drawn from the uniform distribution (0–1 kb), and motif co-localization was evaluated. Mean values over 10 000 simulations are reported in the bar graph, and the P-values of experimental versus random breaks were computed as the frequency of random events as extreme as experimental results. Motif targeting was significantly different (P = 0.0003, Fisher’sexact test) under basal and inflammatory conditions, as pristane-treated mice lost AGCT targeting in favor of break induction outside of canonical AID hotspots. (c) Break point localization and motif targeting of Sμ–σδ junctions along the σδ region (up to 1 kb from the 3′ primer). Breaks detected in the σδ region did not show a significant AID signature, with most of the break points detected outside AID motifs. Inflammatory conditions showed no significant differences in AID motif targeting in the σδ region compared with untreated mice (P = 0.43, Fisher’s exact test). pristane to determine whether enhanced the 3′RR super-enhancer. Furthermore, movens of CSR, which cis-regulatory IgH Sμ DNA double-strand breaks affected these results also show that μ–δ CSR is enhancer (if any) may be implicated in the structural profiles of IgD CSR. regulated and that double-strand breaks the control of IgD CSR? Clearly, the 3′RR Among 11 358 929 raw reads, 87 728 for Sμ–σδ CSR are not random breaks. super-enhancer that controls conven- reads were mapped to a switch junction, The structural profiles of IgD junctions tional CSR is not involved in this and the presence of 88 unique Sμ–σδ revealed a majority of junctions with particular IgH recombination. The IgH junctions was documented. As shown micro-homology or insertions compared intronic Eμ enhancer, previously in Figure 1a (right panel), the structural to blunt junctions, that is, a junctional reported to control VDJ recombination profiles of IgD junctions were signifi- pattern that markedly differs from IgG, but not CSR, might be an appropriate cantly affected in pristane-treated 3′RR- IgA and IgE CSR. Such junctional bias is candidate. Understanding the nature of deficient mice revealing less blunt strongly reminiscent of the junctions the transcriptional enhancer that acti- junctions and elevated junctions with produced by the alternative end-joining vates the regulation of IgD CSR remains micro-homology or insertions. Localiza- (A-EJ) pathway rather than the classical an exciting challenge. tions of break points were qualitatively non-homologous DNA end-joining (C- different in both Sμ (Figure 1b, right NHEJ) pathway used for IgG, IgA and CONFLICT OF INTEREST fl panel) and σδ (Figure 1c, right panel) in IgE CSR.9 These results are consistent The authors declare no con ict of interest. pristane-treated mice compared with with the elevated IgD CSR in 53BP1- ACKNOWLEDGEMENTS untreated mice. deficient mice,5 since in the absence of This work was financially supported through A summary of these 190 independent 53BP1, AID-dependent DNA double- a grant from the ANR (project EpiSwitch-3′ junctions provided a description of the strand breaks are repaired via the A-EJ RR 2016). NG was supported through a grant –σ ′ fi fi Sμ δ CSR sequences in 3 RR-de cient pathway. Finally, in an intense unspeci c from the Association de Spécialisation et fl mice at an unprecedented level. Until in ammatory state (pristane injection), d’Orientation Scientifique (Lebanon) and the recently, only a few Sμ–σδ CSR sequences the phenotype of IgD junctions toward municipality of Khiam (Lebanon). HI was have been reported after cloning and an A-EJ process was reinforced. In con- supported through a grant from CORC (FJA/ subsequent sequencing using Sanger’s clusion, double-strand breaks generated NP 2015-109) and the University of Limoges. method.4 In the present study, the posi- during IgG, IgA and IgE CSR are pre- The authors thank the GénoLim platform tions of IgD junctions (in terms of dominantly repaired through the and the Cytometry platform of the University distance from Sμ and σδ) and their C-NHEJ pathway, whereas for IgD of Limoges for sequencing and cell sorting. structural profiles (blunt, micro- CSR, there is a shift of DNA end repair FB was supported through a grant from the Fondation Partenariale de l’Université de homology or junction with insertions) to the auxiliary A-EJ pathway. At this Limoges and ALURAD. were similar to those previously reported time, the contribution of homologous wt 7 for mice. These results undoubtedly recombination in IgD CSR remains an AUTHOR CONTRIBUTIONS demonstrate that IgD CSR markedly open question. YD and AS designed the study. NG and HI differs from conventional IgG, IgA and Not only trans but also cis molecular performed the experiments. FB and YD IgE CSR, reflectingthefactthatIgDCSR signals that mediate IgD CSR remain analyzed the data. NG, HI, FB, YD and AS is totally independent on the presence of elusive. Since transcription is the primum drafted the manuscript.

Cellular & Molecular Immunology 3'RR independent IgD CSR HIssaouiet al

874

the enigmatic IgD class-switch recombina- 7 Ghazzaui N, Issaoui H, Saintamand A, ’ 1 Chen K, Cerutti A. New insights into the tion via germline deletion of the IgH 3 Boyer F, Denizot Y. Analysis of IgD CSR enigma of immunoglobulin D. Immunol Rev regulatory region. JExpMed2014; 211: junctions by high-throughput sequencing. – 2010; 237:160–179. 975 985. Immunol Lett 2017; 188:86–88. 2 Saintamand A, Rouaud P, Saad F, Rios G, 5 Choi JH, Wang KW, Zhang D, Zhan X, 8 Vincent-Fabert C, Fiancette R, Pinaud E, Cogné M, Denizot Y. Elucidation of IgH 3’ Wang T, Bu CH et al. IgD class switching is Trufﬁnet V, Cogné N, Cogné M et al. Geno- region regulatory role during class switch initiated by microbiota and limited to mic deletion of the whole IgH 3’ regulatory mucosa-associated lymphoid tissue in mice. recombination via germline deletion. Nat region (hs3a, hs1,2, hs3b, hs4) dramatically Proc Natl Acad Sci USA 2017; 114: Commun 2015; 6:7084. affects class switch recombination and Ig 3 ChenK,XuW,WilsonM,HeB,MillerNW, E1196–E1204. 6 Boyer F, Boutouil H, Dalloul I, Dalloul Z, secretion to all isotypes. Blood 2010; 116: Bengtén E et al. Immunoglobulin D – enhances immune surveillance by activating Cook-Moreau J, Aldigier JC et al. CSReport: a 1895 1898. antimicrobial, proinﬂammatory and B cell- new computational tool designed for auto- 9 Chang HHY, Pannunzio NR, Adachi N, stimulating programs in basophils. Nat matic analysis of class switch recombination Lieber MR. Non-homologous DNA end join- Immunol 2009; 10:889–898. junctions sequenced by high-throughput ing and alternative pathways to double- 4 Rouaud P, Saintamand A, Saad F, Carrion C, sequencing. JImmunol2017; 198: strand break repair. Nat Rev Mol Cell Biol Lecardeur S, Cogné M et al. Elucidation of 4148–4155. 2017; 18:495–506.

Cellular & Molecular Immunology